OpenAI Wants to Stop Deepfakes From Overwhelming Child Abuse Tiplines

Posts from this topic will be added to your daily email digest and your homepage feed.

OpenAI Wants to Stop Deepfakes From Overwhelming Child Abuse Tiplines

TL;DR

  • OpenAI's new classifier improves the detection of abusive imagery, including AI-generated content.
  • The tool reduces false positives for human review by 48% and doubles the detection rate of abusive material.
  • It offers free access to vetted organizations and provides guidance for developers and training for law enforcement.
  • The increase in AI-generated CSAM reports poses a significant challenge to organizations like the NCMEC CyberTipline.
  • OpenAI's ChatGPT image generator now includes warnings and pauses for users attempting to create CSAM, directing them to support resources.
  • Laws in many states now cover synthetic CSAM, but collaboration with tech companies is a novel approach to prevention.